Membrane insertion of gap junction connexins: polytopic channel forming membrane proteins

The results presented in this report provide direct evidence of the cotranslational membrane integration of connexin polypeptides into the ER membrane. The integration occurred in an SRP-dependent manner. The membrane integration was accompanied by an unprecedented and efficient proteolytic processing of the connexins when translocated in vitro into ER-derived microsomes, removing an aminoterminal portion of all connexins analyzed. Furthermore, the microsomes used in this analysis contained endogenous uncleaved full-length /5~ connexin, indicating that the pancreatic microsomes generated from the ER during isolation were competent to integrate connexins correctly when the cells were intact. An analysis of connexins generated in beterologous eucaryotic expression systems indicated that this proteolytic processing occurred also under in vivo c~nditions of elevated connexin expression when connexin polypeptides were translocated into the ER membrane. The expression of ot~ and/31 GJ protein consistently resulted in the generation of variable amounts of processed cormexin polypeptides ( Fig. 8 A) , independent of the expression system. Subcellular fractionation provided evidence that the ER membranes contained both the processed and full-length connexin polypeptides, while the Golgi membranes and the PM contained only full-length connexins (Fig. 8 B) . BHK cells expressing different amounts of connexin protein also showed that the generation of the processed polypeptides correlated with the level of protein expression (Fig. 8 A) . GJ protein cDNAs were cloned into eucaryotic protein expression vectors and expressed in yeast, baculovirus, or stable transfected BHK cells, respectively. Yeast proteins were subjected to SDS-PAGE and transferred to nitrocellulose membranes for immunoblot analysis with the cormexin-specific anti-peptide antibodies GAP 10, ill J, and ~/IS. Nontransfected yeast cell lysates were analyzed as controls. Cormexins expressed in the presence of [35S]methionine in baeulovirus-infected cells and BHK cells, respectively, were immtmoprecipitated from complete cell lysates using the anti-peptide antibodies Bt 1-6, cqJ, c~lS (at GJ protein), or BtB, BiJ and BtS (Bi GJ protein), resolved by SDS-PAGE, and visualized by fluorography. Synthetic cormexin RNAs translated in the absence (-) or presence (+) of microsomes were analyzed in parallel as controls. In all three systems, the NH2-terminally processed, faster migrating cormexins (cC and /~13 were detected together with the unprocessed full-size connexins (c~1 and/~l). The BHK cell expression system allows the expression of variable amounts of connexin protein, since the connexin eDNA is inserted behind a heavy metal inducible promoter. Increasing the induction conditions (0.05, 0.1, 0.125, and 0.15 mM zinc) resulted in an equal increase in connexin expression for both full-size and NH2-terminally processed connexins (lanes 14-19 and 22-30) . (B) Highly induced BHK cells expressing c~ or ~1 GJ protein were separated into subeellular #actions containing either rER membranes, Golgi membranes, or PM, respectively, as described in Materials and Methods. Subcellular fractions were subjected to SDS-PAGE and transferred to nitrocellulose membranes for immunoblot analysis with GJ protein-specific anti-peptide antibodies GAP 10,/3tS and cqS. Complete cell lysates were analyzed in parallel as controls. Full-size (at,/~) and NH2-terminally processed GJ proteins (~', ~t') were detected in t